Hi Jecksons,



First I want to make a couple of points about vernacular. By "heap" I believe you mean the "stackup," or the stack of layers that make up the board. Also, by "zigzag" I believe you mean "diagonal." In other words, I believe you are talking about rotating the coordinate plane of one side of the board, or the components on one side, so that the components on that side are not orthogonal to the other side.

Also, by "up" and "bottom" I understand you to mean "top" and "bottom," as in the top and bottom layers of the board. It's perfectly acceptable to refer to the two outer layers of a circuit board as top and bottom, although often you will find all of the layers referred to by number (i.e. layer 1, layer 2, layer 3, ..., layer n). In this case layer 1 is the top layer and layer n is the bottom layer. However, since many boards now incorporate components on both sides, you will never see the outer layers referred to as the "parts" layer or the "trace" layer anymore.

If I am wrong about any of these assumptions let me know. For now I will proceed with these understandings.

You say that rotating one side off-orthogonal will "make an easy task." In fact it will not, and what's worse it can actually mess up your autorouter. Look up and study "channels," or "routing grid" and "manhattan distances" in your software's manual to learn why.

As for why you believe it will make your task easier, I'm guessing it's about visualizing, or visually distinguishing, what you are working on. Use colors instead.

All EDA software uses varied colors to display the different layers, specifically so the different layers can be distinguished by the designer. The software will automatically display the color of whichever layer is active on top so that no details are obscured by features on other layers.

You also need to understand that placing a component on the bottom of the board means it must be flipped (mirrored). Again, all EDA software does this automatically when you place the component on the bottom layer. Somewhere in your software is a button or window that will let you specify which layer a component is placed on. It's probably in a properties dialog box for the part in question. Or maybe it's somewhere else. In any case, find and use this feature.



Depending on the particular instance, this might be a very clever and elgant design solution, or it could just be a lazy and (hopefully) lucky designer.

Some chips, like memory chips, are designed with pinouts that are amenable to this type of interconnection. You can place chips directly onto the same pads, with one chip in mirror image, and they both work. In such cases as this it is an elegant technique that allows for bus net management.

More likely however is that a designer just placed opposing SMT parts in the same coordinates. He figured that since the solder pads exist only on their respective planes, they remain galvenically isolated (they don't short out to each other) and all is well. He probably thought it made his board look nice and neat all organized in pretty ranks and files.

He's kind of like the designer who thought his board looked nice with the SIP headers lined up across the middle of his board (probably where he split a PWR/GND plane) without ever thinking about the effect of a perforation line. Picture the perforation line on a roll of paper towels.

It's true that SMT pads on opposite sides of the board won't short together, but they will capacitively couple. If the circuit still works, it's likely because the designer was lucky. Unless you have a specific reason to do otherwise, it's a good idea to shift parts so they don't align. Or if you need to line them up, at least do the math to make sure that the added capacitive coupling isn't going to mess anything up.

I hope this helps,

Joe